What is deep tech, and why does nobody tell you the selling is the hard part

At a deeptech event in Delft I asked a room of founders what they sold. One said photonic chips. One said enzymatic recycling. One said, and I quote, a novel approach to thermal management. Then I asked who bought it, and the room went quiet in a specific way. Not because they did not know their customer. Because the honest answer was: we are still figuring that part out.

That quiet is what this article is about. Deep tech is easy to define on a slide and surprisingly hard to sell, and the gap between those two things is where most of these companies live.

Deep tech is a funding category, not a customer category

The definition you hear: companies built on real scientific or engineering advances. Long development times, hard-to-copy technology, physical products or deep infrastructure. Fine. But notice what that definition is for. It is for investors and policymakers. It describes how hard the thing was to build.

Your customer does not care. A plant manager does not buy deep tech. She buys fewer unplanned stops, lower energy per unit, a process that passes the audit. The moment you understand that deep tech describes you and not the purchase, a lot of confusing sales behaviour stops being confusing.

The technology is rarely the reason deals die

Founders assume deals are lost on technical merit. Sometimes. Far more often the deal dies in a process nobody explained to them: the champion who could not get budget authority, the capex committee that needed a payback case nobody built, the supplier qualification that started after the yes and quietly ate a year.

Selling software is about being chosen. Selling deep tech is about surviving: surviving the evaluation, the qualification, the internal meeting you are not in, the eighteen months between the handshake and the invoice. That is a different sport, and most founders trained for the first one.

The buyers are committees, and the committees are scared

In consumer markets a customer risks twenty euros. In enterprise software, a licence and some embarrassment. In deep tech, the person who signs off on your product risks the line, the plant, and sometimes the licence to operate. An engineer who champions a new material that fails in production does not get a bad quarter. She gets a career event.

So industrial buying is designed to spread risk across many shoulders, which means no single person can say yes, while almost anyone can say no. Selling into that is not persuasion. It is de-risking, person by person, on paper, over quarters.

The clock runs on years, and nobody budgeted for that

A software startup can learn in weeks. Ship, measure, change the pitch. A deeptech company gets a handful of real commercial attempts per year, because each cycle, from first meeting to a decision that counts, runs six months to two years. Learning is slow, expensive, and unforgiving.

This is the part nobody tells you at the demo day: the burn rate of a deeptech startup is not engineering salaries. It is the commercial timeline. Every pricing mistake, every wrong beachhead, every pilot with no defined next step costs a season, not a sprint. Which is why the pilot agreement matters more than the pitch deck.

So what do you do with this

Three things follow. First, sell the problem's cost, not the technology's elegance. Second, learn the buying process of your specific industry before you build the sales pitch, because the process is the pitch. Third, treat commercial capability as seriously as the lab: it takes the same years to build, and it fails just as expensively.

That third one is the reason SLSCRW exists. Deeptech companies should not have to hire a full commercial team before they have the commercial activity to justify one. We build the commercial engine around your technology, alongside you, until there is enough revenue to justify your own. If the quiet in that Delft room sounded familiar, tell us what you have built.